Wilson: Benching Ike

A week before the season the Steelers gave cornerback Ike Taylor a 5-year, $22.5 million contract extension and 10 weeks later he's been benched. But maybe it's the right thing to do. Blogger Ryan Wilson digs through the data and unearths a few surprises.

In Chris Rock's 1996 "Bring the Pain" he had a bit about the O.J. Simpson murder trial. Rock talked about how O.J.'s ex-wife, Nicole Brown-Simpson, let her "friend" Ron Goldman drive around L.A. in an O.J.-purchased luxury vehicle. And how Simpson, after seeing Goldman sporting around town in the paid for O.J.-mobile, might've gotten a little upset. After setting up the story, Rock delivered the punch line: "I'm not saying he shoulda killed her … but I understand."

When I found out Bill Cowher might bench Ike Taylor for Bryant McFadden this Sunday, I was reminded of that Chris Rock line. I'm not saying Cowher shoulda benched Ike ... but I understand.

It's not like Taylor has suddenly morphed into DeWayne Washington. In fact, before last week's thumpin' at the hands of Javon Walker, nobody had mentioned Taylor's name in four or five weeks. And that's a good thing when you're an NFL cornerback. Maybe some of that has to do with all the other problems on the team, but if he was really stinking up the joint, you think somebody would've noticed before now. Or maybe not.

Taylor's not without his problems, though. Early in his career he struggled with tackling and we all know he can't catch; he even admitted to dropping no fewer than nine picks last season. But after spending the early part of the 2004 season in Bill Cowher's doghouse (Taylor couldn't beat out Willie Williams for an active game day roster spot, for cripes sake), he's shown steady improvement the past two-plus seasons. And up until last Sunday's disaster against the Broncos, I would've probably said his development had been monotonically increasing. So much for that. And Cowher thought so little of Taylor's most recent performance it'll probably cost the fourth-year cornerback his starting job this week. It's also worth noting that the Pittsburgh Steelers just re-signed Taylor to a 5-year, $22.5 million contract. Two months ago.

But is it just the Broncos game that led to Taylor's demotion or has it been a decision weeks in the making, and because of everything from a concussed quarterback to mouthy linebackers, it's gotten lost in the media mix?

Put differently, is it the right move? Well, that depends on how you define "right." If you mean that Taylor has been the only player to struggle this season and deserves to be benched then, obviously, it's very, very wrong. At various points in 2006 Ben Roethlisberger, Cedrick Wilson, Max Starks, Kendall Simmons and the entire special teams have all been more deserving of some pine time than Taylor.

If, instead, you mean Taylor hasn't played like a No. 1 cornerback – or at least how the team envisions a No. 1 cornerback should play – then that's a different issue. Luckily, we can look at the data to get some insights into Taylor's 2006 season.

Using NFL.com's play-by-play data for the Steelers' eight games, I charted all passes thrown to receivers covered by one of three Pittsburgh cornerbacks: Taylor, McFadden and Deshea Townsend (Ricardo Colclough had so few attempts it wasn't worth including him). Keep in mind the data aren't perfect; for example, just because Taylor tackles Marty Booker after a 15-yard reception, it doesn't mean Taylor was responsible for Booker. It could be that the Steelers were in zone coverage, Taylor was responsible for the flat, Booker settled in the middle of the field, broke a tackle, and Taylor came from his area of the field to make the play. There are also a few instances when the play-by-play data don't identify the defensive player in coverage on a pass play. With those caveats, here's what we've got through nine weeks of some pretty awful football:

Table 1.

Name       totpass  incompl  compl  yds   comp%
Taylor       50       18       32   442   64%
Townsend     32       11       21   175   66%
McFadden     24       12       12   175   50%
*Taylor      41       14       27   317   66%
Table 2.
Name        WR1  WR2  WR3  TE  RB
Taylor      60%  14%  12%  10%  4%
Townsend     6%  35%  26%  19% 13%
McFadden    17%  13%  38%  21% 13%
*Taylor     51%  17%  15%  12%  5%
Table 3.
Name       yds/att  Wyds/att  yds/comp  Wyds/comp  +Plays
Taylor       8.84   11.13      13.81     17.38       47%
Townsend     5.47    7.89       8.33     12.02       62%
McFadden     7.17    9.35      14.33     18.69       57%
*Taylor      7.73    9.00      11.74     14.06       45%
A few things before discussing the findings: Table 1 looks at total passes thrown in the direction of the cornerback in question, incompletions, completions and yards allowed; Table 2 lists what percentage of passes were to certain types of receivers, by cornerback; Table 3 breaks out average yards per attempt (unweighted and weighted), average yards per completion (unweighted and weighted), and the percentage of all plays that were judged to be positive plays for the defense. (For simplicity, I just looked at whether the play gained enough yardage for a first down or touchdown. If not, it was a positive play for the defense; otherwise, it wasn't.)

For the weighted measures, I assigned weights based on the number of receptions made by the opposing team's No. 1 receiver, No. 2 receiver, No. 3 receiver, TE and RB, by cornerback. Because Taylor primarily faces No. 1 receivers, you would expect his yards per attempt average to be higher than if he just covered running backs (on the assumption that No. 1 receivers are better pass catchers, in general, than running backs). Using this weighted measures framework, we can compare across cornerbacks even though Taylor faced No. 1 receivers 60 percent of the time, McFadden 17 percent of the time and Townsend only six percent of the time.

Finally, the last row in each column (denoted by a "*Taylor") shows what Taylor's numbers look like if we throw out the Broncos debacle.

In looking at the numbers, the first thing that jumps out is how frequently teams are throwing in Taylor's direction. This is a little misleading, however. Teams have tried to throw deep (15 yards or more down the field) on Taylor 15 times and have been successful six times – two coming last week. That's a 40 percent success rate on deep passes, which, without context, still sounds pretty bad.

Overall, offenses are completing 64 percent of passes thrown in Taylor's direction. Again, this sounds high – especially when you consider Hines Ward is only catching 59 percent of passes thrown to him – but as has been the case for most of 2005 and 2006, teams are also dinking and dunking Taylor more than they're throwing deep down the field. In fact, Taylor has been in on 20 plays identified as "short passes" in the play-by-play data and 17 have gone for either first downs or touchdowns. That's a success rate of 85 percent, which is rarefied Chad Scott territory.

And while opposing teams have admitted to game planning for Townsend, he's faired slightly better than Taylor. Teams have thrown deep six times against him, and were successful on three occasions; on short passes, offenses were 7-for-21 (31%), which is certainly favorable when compared to Ocho Cinco above.

McFadden is slightly better than Townsend; he's 3-for-6 on deep passes and 4-for-14 on short passes (29%).

Table 3 offers more evidence that Taylor hasn't been playing like a top-flight cornerback – or maybe even the best cornerback on his team – even after adjusting for the receiver. Townsend and McFadden both sport lower yards allowed per attempt, and make more positive plays.

When I first learned of Taylor's demotion I didn't think anything more of the move than a desperate head coach trying to spark his 2-6 team out of a nine-week coma. After seeing the data, however, I think the move is probably justified. Again, benching Taylor isn't going to magically solve Pittsburgh's problems with turnovers, or the offensive line's inability to run-block, but if McFadden and Townsend make the Steelers' secondary better – at least in the short run – then why not give them a chance? And it's not like Taylor will be "Duced" and not dress for the game – I'm sure he'll be in the sub packages and on special teams.

Despite these revelations – and I was pretty shocked by the data – I still think Taylor is a good NFL cornerback and will only get better. And whenever I hear some variation of the following:

"…Many always talk about Ike shutting down Chad Johnson. But he also gets torched more than his talent should allow. Marvin Harrison beat him deep by a good 5 yards on the Colt's opening drive last year on Monday night. Seattle went after Ike repeatedly at the beginning of the Super Bowl last year. Of course he got abused last week by Walker…"
I give my stock response:
Pittsburgh plays a lot of Cover-3 behind all the zone blitzing Dick LeBeau likes to do, and one of the cornerback's responsibilities is to keep the play in front of them. That, in part, explains (a) Taylor giving big cushions, and (b) why he gives up so many short completions.

So sure, a lot of teams throw at him, but a lot of it is of the underneath variety, and in Pittsburgh's scheme, underneath is better than deep, especially when cornerbacks are often in single-coverage. Another reason teams might throw at Taylor is because he can't catch. So even on a deep ball, from the offense's perspective, the worst thing to happen is an incompletion.

Still, it's hard to argue that Taylor has been anything other than ordinary this season. But like I said, Ike will be fine.

In the meantime, what's the worst that can happen by starting McFadden and Townsend, a loss? Did you see the Oakland game? This'll be a walk in the park.

Steel City Insider Top Stories